Improvement of APACHE II score system for disease severity based on XGBoost algorithm

Table 5 Performance of the XGBoost model with different variables

Variables	Number of variables	Accuracy [95% Cl]	Precision [95% Cl]	Recall [95% Cl]	F1 [95% Cl]	AUC [95% Cl]
Urine	15	0.862 ± 0.005 [0.855, 0.869]	0.839 ± 0.010 [0.825, 0.853]	0.862 ± 0.005 [0.855, 0.869]	0.832 ± 0.008 [0.822, 0.841]	0.788 ± 0.006 [0.780, 0.796]
Bilirubin	16	0.865 ± 0.004 [0.860, 0.871]	0.845 ± 0.007 [0.836, 0.855]	0.865 ± 0.004 [0.860, 0.871]	0.839 ± 0.005 [0.832, 0.846]	0.802 ± 0.006 [0.793, 0.810]
Albumin	17	0.864 ± 0.006 [0.856, 0.872]	0.843 ± 0.010 [0.829, 0.857]	0.864 ± 0.006 [0.856, 0.872]	0.838 ± 0.008 [0.827, 0.850]	0.805 ± 0.006 [0.797, 0.814]
BUN	18	0.866 ± 0.004 [0.861, 0.872]	0.847 ± 0.007 [0.837, 0.856]	0.866 ± 0.004 [0.861, 0.872]	0.842 ± 0.006 [0.834, 0.850]	0.811 ± 0.005 [0.804, 0.818]
Glu	19	0.867 ± 0.004 [0.860, 0.872]	0.846 ± 0.007 [0.836, 0.856]	0.867 ± 0.004 [0.860, 0.872]	0.841 ± 0.006 [0.832, 0.850]	0.811 ± 0.004 [0.805, 0.816]

ISSN: 1472-6947